A reconfigurable fabric for accelerating large-scale datacenter services

TLDR

Datacenter workloads demand high compute, flexibility, power efficiency, and low cost, yet improving all simultaneously is difficult. The study designs and evaluates a composable, reconfigurable fabric that accelerates large‑scale datacenter services beyond commodity servers. The fabric is a 6×8 torus of Stratix V FPGAs, one per server, linked via PCIe and 10 Gb SAS cables, deployed across 1,632 servers to accelerate Bing search ranking with robust engineering for failure resilience. Under high load, the fabric increases per‑server ranking throughput by 95% at fixed latency or reduces tail latency by 29% while maintaining throughput.

Abstract

Datacenter workloads demand high computational capabilities, flexibility, power efficiency, and low cost. It is challenging to improve all of these factors simultaneously. To advance datacenter capabilities beyond what commodity server designs can provide, we have designed and built a composable, reconfigurablefabric to accelerate portions of large-scale software services. Each instantiation of the fabric consists of a 6x8 2-D torus of high-end Stratix V FPGAs embedded into a half-rack of 48 machines. One FPGA is placed into each server, accessible through PCIe, and wired directly to other FPGAs with pairs of 10 Gb SAS cables In this paper, we describe a medium-scale deployment of this fabric on a bed of 1,632 servers, and measure its efficacy in accelerating the Bing web search engine. We describe the requirements and architecture of the system, detail the critical engineering challenges and solutions needed to make the system robust in the presence of failures, and measure the performance, power, and resilience of the system when ranking candidate documents. Under high load, the largescale reconfigurable fabric improves the ranking throughput of each server by a factor of 95% for a fixed latency distribution--- or, while maintaining equivalent throughput, reduces the tail latency by 29%